Schema and Structured Data · Guide

What Is Video Schema Markup?

The one type in this section becoming more relevant rather than less. A video cannot be read, so everything anything knows about its contents comes from what surrounds it. Most businesses publishing video describe none of it.

Updated: August 2026
Written by: Andrew Odgers, Managing Director
Reading time: 9 minutes
The claims it makes

What It Asserts

Title, description, thumbnail, duration, upload date and where the file lives. That list exists because a video is opaque: nothing reading a page can watch it, so the description is the only account of what it contains.

Title and description. What it is about.

The subject and a summary of the contents. These carry more weight than on other types, since there is no text in the video itself to fall back on.

Thumbnail. The still image.

What appears before somebody presses play. Block four covers why this deserves more attention than it usually gets.

Duration. How long it runs.

Stated because somebody deciding whether to watch is answering a question about their time. Short and specific rather than approximate.

Upload date. When it appeared.

Useful for anything where currency matters, being the property that quietly ages a video that was never replaced.

Where the file lives. The location.

The address of the video itself rather than of the page displaying it. Block three covers why that distinction causes so much confusion.

Our own observation, stated as such

Why It Is Becoming More Relevant

Across our own research for several trade guides, video appears against a substantial share of the demand in some sectors. That is our finding rather than a published figure. It is not something we quantify.

Where we saw it most. Personal training.

Building the guides for that trade, video turned up against a large share of what people search. That was sharper than in any other sector we examined and it changed what we recommended.

Why that sector. People are choosing a person.

Somebody selecting a trainer wants to see how they teach and whether they would tolerate them for an hour. No description answers that and thirty seconds of footage does.

Where the same logic applies. Anywhere judgement is personal.

Therapists, instructors, coaches and anybody whose manner is part of the service. The pattern is choosing a person rather than comparing a product.

What follows for those businesses. Describing it matters.

A business publishing video without describing it has produced something nothing can read. The effort of making it is already spent. The description costs a fraction of that.

Why we state this cautiously. It is our own work.

We looked at particular sectors at a particular time. Publishing a percentage would imply a precision the observation does not have. The direction is the useful part.

The distinction that causes confusion

Hosted Or Embedded

A video held on a platform is already described there. One embedded on your own page may need describing by you. The description should point at the actual file rather than at the page showing it.

What hosting means. Somebody else holds it.

The video lives on a platform, which describes it thoroughly because that is what platforms do. Their description exists whether or not you add one.

What embedding means. Displayed here, held there.

Your page shows a player while the file itself sits elsewhere. The page is a place the video appears rather than where it lives.

Why that matters for the description. Two things are being confused.

People describe the page rather than the video, which produces a description pointing at itself. The location property should give the file, not the address you are already on.

When to describe an embedded video. When the page is the point.

Where somebody arriving on your page is the outcome you want, describing it there makes sense. Where the platform is where you want people, its own description already serves.

The duplication question. Same as elsewhere.

Both describing the same video is the pattern the platform page warns about. Decide which is authoritative rather than maintaining two loosely.

Two properties worth attention

Thumbnails And Duration Matter

These two can appear where a video is presented, which makes them unusual among properties nobody thinks about. A missing or poor thumbnail wastes whatever the video cost to make.

What the thumbnail does. Decides whether anybody watches.

It is the entire visible representation of the video before somebody commits. A frame grabbed automatically is frequently somebody mid-blink, which is not the impression intended.

What a considered one looks like. Chosen rather than sampled.

A deliberate still that shows what the video is about and reads clearly at a small size. That takes a minute and outlasts everything else about the description.

What duration does. Answers the time question.

Somebody deciding whether to watch wants to know what they are committing to. A stated length lets them decide rather than guess. Short videos benefit most from saying so.

Why these two get neglected. They feel cosmetic.

Both look like presentation rather than substance, so they get left to whatever the platform generated. They are the properties most likely to be seen.

What accuracy requires. The obvious.

The stated duration should be the actual duration and the thumbnail should come from the video. Both are assertions like any other here.

The underused half

Transcripts And Captions

Text accompanying a video is what makes its content findable. That sits alongside the markup rather than inside it. Most businesses skip it entirely.

Why text is the mechanism. The video is opaque.

The markup describes a video in a sentence. A transcript contains everything said in it, which is a different order of information about what the thing actually covers.

What that produces. A page with substance.

A video with a transcript beneath it is a page containing an argument rather than a page containing a player. The second has almost nothing on it.

Why businesses skip it. It looks like duplication.

Publishing the words alongside the video feels redundant when the video says them. It is the only version anything reading the page can access.

The other reason it is worth doing. People read.

A proportion of your audience would rather skim text than watch. Some cannot watch where they are. Captions and transcripts serve them as well as serving discovery.

The practical order. Transcript first, markup second.

If you only do one, publish the text. The description helps a video be understood and the transcript is what there is to understand.

The assertion rule, in its commonest leftover form

Do Not Mark Up A Video That Is Not There

Markup describing a video removed from the page is a false assertion. This type produces that leftover more often than most, because videos disappear for reasons nobody connects to the description.

How videos vanish. Four ways.

Replaced with a newer one, removed during a redesign, taken down on the platform holding it, else an account closing and taking its contents with it.

Why the description survives. Nothing links them.

Removing a player from a page does not remove the description of it. The two are separate, which is the advantage of the separated format working against you.

What the site then claims. A video exists here.

The page asserts a video with a title, a length and a location while there is nothing on the page at all. That is describing a page that does not exist.

Why this one is easy to miss. The page still reads fine.

A page that lost its video usually still has its text and looks complete. Nobody visiting notices anything absent, so nothing prompts the correction.

When to check. Whenever a video changes.

Replacing or removing a video should include updating what the page says about it, in the same job rather than a later one.

Practical scoping

Where It Genuinely Helps

Businesses where somebody is choosing a person, plus businesses demonstrating something that has to be seen. Outside those, video is an expensive way to say what a paragraph would say.

Choosing a person. The strongest case.

Trainers, therapists, instructors and consultants, where manner is part of the service. Our personal trainers material covers what we found there. The recommendation follows from it.

Demonstrating something visual. The second case.

Where a process, a finish or a result is genuinely hard to convey in words. A before and after, a technique, a scale that only reads visually.

Where it adds little. Explaining a service.

A video of somebody describing what a company does is a paragraph read aloud, more slowly and less searchably. That is the commonest business video and the least useful.

The cost worth weighing. Production and maintenance.

Video is expensive to make and awkward to update, since a change means reshooting rather than editing. A page that changes annually is a poor candidate.

What to do before committing. Ask what it shows.

If the answer is something text could carry, write it instead. Google SERP features explained covers where video appears. The full series is on the schema and structured data guide.

Website migrations

Publish the
transcript before
the markup.

The description accounts for a video in a sentence. A transcript contains everything said in it, which is a different order of information about what the thing covers. It also serves the people who would rather skim than watch.

On every page we build:

Types that actually apply Transcripts alongside video Deliberate thumbnails One authoritative source Article description Questions and answers Breadcrumb structure Recheck after changes

If a video would only say what a paragraph says, we will tell you to write the paragraph.

The full guide series

Every guide.
One practice.

What structured data is, which types apply to your business, how to test it, where the results features come from and what to do when nothing appears.

Questions people ask

Video Markup, Briefly

What does video markup describe?
Title, description, thumbnail, duration, upload date and where the file lives. That list exists because a video is opaque: nothing reading a page can watch it, so the description is the only account of what it contains. The title and description carry more weight here than on other types, since there is no text to fall back on.
Is video worth bothering with for our sector?
It depends whether customers are choosing a person. Across our own research for several trade guides, video appeared against a substantial share of demand in some sectors, most sharply for personal training, where somebody wants to see how a trainer teaches. We do not quantify that, since it was our own look at particular sectors at a particular time.
Our video is on a platform. Do we describe it too?
Often not. A hosted video is already described thoroughly where it lives, whether or not you add anything. Describing it on your page makes sense where somebody arriving on your page is the outcome you want. Doing both creates two descriptions of one video, so decide which is authoritative rather than maintaining two loosely.
Do thumbnails really matter?
More than anything else here, because the thumbnail is the entire visible representation of the video before somebody commits. A frame grabbed automatically is frequently somebody mid-blink. A deliberate still that shows the subject and reads clearly at small size takes a minute and outlasts everything else about the description.
Should we publish a transcript?
If you only do one thing, do that. The markup describes a video in a sentence while a transcript contains everything said in it, which is a different order of information. It also serves the proportion of your audience who would rather skim than watch, else cannot watch where they are. A video with a transcript is a page containing an argument rather than a player.
We replaced a video. Does anything need doing?
Yes. This is also the most frequent leftover with this type. Removing a player does not remove the description of it, so the page asserts a video with a title, a length and a location while containing nothing. It is easy to miss because the page still reads fine and nobody visiting notices anything absent.